Revise usage of `encoding` in more CompilationUnitResolver methods

Quoting @jukzi on https://github.com/eclipse-jdt/eclipse.jdt.core/pull/2560#discussion_r1654179571
> Making the encoding part of the API seems wrong as currently the encoding is not used consistently: Only Batch Compiler uses it in org.eclipse.jdt.internal.compiler.batch.CompilationUnit.getContents() while the IDE uses org.eclipse.jdt.internal.core.util.Util.getResourceContentsAsCharArray(IFile) which disrespect the given classpath encoding but uses the encoding configured in workspace and the utf bom.

We should revise usage of `encoding` here to either actually use it, or get rid of it.